#archlinux32 | Logs for 2022-09-02

[00:54:26] -!- deep42thought has joined #archlinux32
[00:58:38] -!- oaken-so1rce has joined #archlinux32
[00:58:57] -!- girls has quit [Ping timeout: 244 seconds]
[00:58:58] -!- buildmaster has quit [Ping timeout: 244 seconds]
[00:58:58] -!- oaken-source has quit [Ping timeout: 244 seconds]
[01:11:36] -!- buildmaster has joined #archlinux32
[02:06:07] <bill-auger> website is not responding "504 Gateway Time-out"
[05:23:31] -!- ejjdhfjsu has joined #archlinux32
[07:50:55] -!- titus_livius has joined #archlinux32
[07:53:44] -!- buildmaster has quit [Remote host closed the connection]
[07:56:23] -!- buildmaster has joined #archlinux32
[08:02:41] -!- ejjdhfjsu_ has joined #archlinux32
[08:04:59] -!- ejjdhfjsu has quit [Ping timeout: 244 seconds]
[08:53:42] -!- ejjdhfjsu__ has joined #archlinux32
[08:56:47] -!- ejjdhfjsu_ has quit [Ping timeout: 268 seconds]
[11:05:40] -!- ejjdhfjsu_ has joined #archlinux32
[11:08:15] -!- ejjdhfjsu__ has quit [Ping timeout: 252 seconds]
[11:20:34] -!- titus_livius has joined #archlinux32
[11:20:49] -!- abaumann has joined #archlinux32
[11:20:50] <buildmaster> Hi abaumann!
[11:20:53] <buildmaster> !rq abaumann
[11:20:53] <phrik> buildmaster: <abaumann> I'm using till until proven wrong.. ;-)
[11:21:03] <abaumann> bill-auger: yeah, sorry, archlinux32.org is running amok..
[11:21:37] <abaumann> mysql was heavy on IO/wait
[11:21:41] <abaumann> not really sure, why.
[11:23:30] -!- abaumann has quit [Client Quit]
[11:35:04] <deep42thought> reason is probably just "because mysql"
[11:36:52] <Foxboron> deep42thought: how does archlinux32 currently build packages?
[11:37:34] <deep42thought> it uses a fork of the devtools from archlinux and a couple bash scripts
[11:37:45] <Foxboron> https://buildbot.pkgbuild.com
[11:37:45] <phrik> Title: Buildbot (at buildbot.pkgbuild.com)
[11:37:50] <Foxboron> I'm working a bit on build automation for Arch
[11:38:11] <Foxboron> I plan on announcing it soon the project soon, but I'm wondering if we should include builders for ports if they are available
[11:38:38] <deep42thought> does it some kind of dependency resolution (i.e. orders the builds by dependencies)?
[11:38:59] <Foxboron> Not yet
[11:39:28] <deep42thought> because that's (in my opinion) the biggest task of the shell scripts that we have around devtools32
[11:40:15] <deep42thought> if buildbot has at least a basic level of dependency conflict avoiding, then we could switch to buildbot for archlinux32, probably
[11:40:45] <deep42thought> often enough, the dependency resolution in our build scripts fails anyway, and we have to resolve manually ;)
[11:41:16] <Foxboron> Where are the scripts?
[11:41:37] <deep42thought> https://git.archlinux32.org
[11:41:38] <phrik> Title: builder - Archlinux32 build system (at git.archlinux32.org)
[11:41:40] <Foxboron> Currently I'm just aiming for rebuilds of v3. But I think having one common build infra for all ports would be nice
[11:42:06] <deep42thought> yeah, that's something, I have dreamed of before, too :)
[11:42:21] <Foxboron> https://md.archlinux.org
[11:42:23] <phrik> Title: HedgeDoc - Collaborative markdown notes (at md.archlinux.org)
[11:42:25] <deep42thought> one build infrastructure to build for x86_64, i686, armv7h, etc.
[11:42:26] <Foxboron> some mockups of frontend tooling
[11:42:30] <Foxboron> https://gitlab.archlinux.org current source
[11:42:31] <phrik> Title: Morten Linderud / archlinux-buildbot ยท GitLab (at gitlab.archlinux.org)
[11:43:17] <deep42thought> the md link redirects me to the main page, I suppose that's an error?
[11:43:22] <Foxboron> mmm
[11:43:30] <Foxboron> https://md.archlinux.org
[11:43:31] <Foxboron> there you go
[11:43:31] <phrik> Title: buildctl - HedgeDoc (at md.archlinux.org)
[11:43:37] <deep42thought> :)
[11:45:19] <Foxboron> Obviously a bit of work until it can be usable. But it's.. something
[11:45:21] <Foxboron> I hope
[11:48:12] <deep42thought> yes, looks promising :)
[11:49:40] <Foxboron> deep42thought: those bash scripts of yours are terrible
[11:49:47] <deep42thought> I know
[11:50:00] <Foxboron> At line 100 i would have reconsidered and done it in something sensible :p
[11:54:41] <Foxboron> deep42thought: So, for the rebuilds. I don't actually think the same problem domain exists.
[11:54:54] <Foxboron> Because in the future the rebuilds that is needed for dependency resolutions would be prebuilt by all archs
[11:54:56] <Foxboron> (in theory)
[11:55:16] <Foxboron> So you don't need to re-resolve the chain. You already know the rebuild that did happen
[11:55:38] <deep42thought> in theory, this is right
[11:55:48] <Foxboron> With the current situation this won't work, obviously
[11:56:19] <deep42thought> well, we could simply build in the same order as you do - but we would also need to build "intermediate" staging-only packages
[11:56:38] <deep42thought> and we would be blocked by every single build failure - no matter how unimportant it actually was ...
[11:57:15] <Foxboron> Sure, but in the future it would be easie to collaborate on the ports
[11:57:59] <deep42thought> yeah, if an additional architecture merely meant adding some more builders to the buildbot, this would be great :)
[11:58:23] <Foxboron> more builders and a git branch on the package repos :)?
[11:58:59] <Foxboron> Then the people building ports would just go "buildctl list builds --arch x86" and you can see failures and the rebuilds not working
[11:59:05] <deep42thought> yes, something like that - what's your source of PKGBUILDs in the buildbot currently?
[11:59:12] <Foxboron> svn and asp
[11:59:38] <deep42thought> I thought, everything was in git nowadays?
[11:59:41] <Foxboron> lolno
[11:59:51] <Foxboron> Still a work in progress
[11:59:58] <deep42thought> ok, I see :)
[12:24:15] -!- abaumann has joined #archlinux32
[12:24:15] <buildmaster> Hi abaumann!
[12:24:15] <buildmaster> !rq abaumann
[12:24:16] <phrik> buildmaster: <abaumann> nono, it's self-cooking :-)
[12:24:27] <abaumann> mysql just performs like sh* on archlinux32.org *grmpf*
[12:24:41] <abaumann> 2x60% CPU, 2x40% system
[12:24:49] <abaumann> I would love to know, what's causing it..
[12:24:56] <deep42thought> Hi abaumann!
[12:25:10] <abaumann> hi deep42thought
[12:34:31] <abaumann> mmh. now it looks ok again.. *shrug*
[12:45:26] -!- abaumann has quit [Quit: leaving]
[13:46:01] -!- ejjdhfjsu_ has quit [Remote host closed the connection]
[14:46:31] -!- drathir_tor has quit [Ping timeout: 258 seconds]
[14:52:57] -!- drathir_tor has joined #archlinux32
[16:10:22] -!- abaumann has joined #archlinux32
[16:10:22] <buildmaster> Hi abaumann!
[16:10:22] <buildmaster> !rq abaumann
[16:10:23] <phrik> buildmaster: <abaumann> Rust will manage to kill linux userland, something Java never managed in 30 years :-)
[16:10:44] <abaumann> ah, sorry guys. My mirror with the chat logs was completely out of sync. Common build infrastructure, sounds great. :-)
[16:11:45] <abaumann> We might need a way to trigger or ignore builds manually, currrently a lot of packages are broken for 32-bit and need special treatement..
[16:12:38] <abaumann> deep42thought: seems Mysql stopped to be Mysql on Archlinux32.org and works ok now.. :-)
[16:15:38] -!- abaumann has quit [Quit: leaving]
[16:41:18] <KillerWasp> maybe the tools need a update and must remove mysql and replace by a plain text output
[16:42:10] <KillerWasp> i don't know if for archlinux must still use the same and they solve with buy a more powerful machine for the work
[17:26:26] -!- GNUtoo has quit [Write error: Connection reset by peer]
[17:27:09] -!- GNUtoo has joined #archlinux32
[17:46:42] -!- T`aZ has quit [Remote host closed the connection]
[19:42:56] <deep42thought> archlinux does not use mysql at all
[19:43:04] <deep42thought> we just introduced it to solve dependency issues
[19:43:08] <deep42thought> (and as a db in general)
[19:43:51] <deep42thought> "plain text output" will not allow to answer questions like "which package can I build, that has all dependencies met, or is in a dependency cycle, or ...?"